For your consideration: Secondary Architectures in Fedora

[Chris Weyl]

On 5/30/07, Jesse Keating <jkeating at redhat.com> wrote:
> On Wednesday 30 May 2007 10:18:55 David Woodhouse wrote:
> > I see no reason to believe that it _would_ "hinder the progress of
> > Fedora on the primary arches". We've seen very little evidence of that
> > so far. As I said, I think you're 'solving' a problem that doesn't
> > exist.
>
> I just simply don't agree with you.  I've _seen_ it hinder the ability to get
> a release out, or a package update out.  I've _seen_ packages sit unbuildable
> for hours if not days until the one person with arch specific knowledge can
> look into whats going on.  Every new package will have to be 'bootstrapped'
> into an arch.  I can't expect new Fedora contributors to be completely
> stalled on getting software into Fedora, into the hands of millions of users
> simply because the software doesn't build for some obscure arch that somebody
> is working on getting going for Fedora, that has at most a couple hundred
> users.

<preface>Ok.  The next paragraphs are intended as constructive
criticizm.</preface>

Or, in a slightly different scenario, have new requirements on
maintainters to support additional arches -- some of whom are either
actively ticked or increasingly concerned about how the recent changes
to building and releasing will impact their workflow.  (I hardly need
to cite the recent threads on -maintainers.)  Changes, I hate to say,
which aren't entirely clear yet as to their operation or impact aside
from "we're going to do this now."

I've been saying this a lot lately, but it seems (IMHO) to be being
obscured with the recent core merge -- our maintainer base is largely
_volunteer_.  I certainly don't find deposits from RedHat, Inc on my
bank statement.[1] We need to balance whatever gain we get from
changes that increase burden or workload against the fact that these
volunteers freely donate their time and skills because they get
something out of it too...  If we push the pain level above the return
they get from contributing, what motivation do they have to contribute
anymore?

Now, back on point, the SecondaryArch proposal as it stands satisfies
a couple good points Jesse was making, but also allows the existing
package base to be built for these other arches just as EPEL allows
the package base to be built for RHEL w/o incurring additional
workload on the maintainer...  I'm sure maintainers will address
trivial bugs with other arches,   but people who care about these
secondary arches can watch for and solve the more serious ones; ones
they may be the only people who have any idea how to resolve.

This isn't about "getting maintainers out of work", this is about
trying to balance two legitimate needs and goals:  leveraging the
extant Fedora infrastructure to build for secondary arches, and
maintaining the interest and commitment of our maintainer base.

Isn't this what we want?  Enable people who care about something to do
it without impacting everyone else?

                                              -Chris

[1]  Note, however, I wouldn't be opposed to it.  Paypal or small,
unmarked bills at the usual location would also be fine ;)
-- 
Chris Weyl
Ex astris, scientia